'
Tasks: TOP => search:reindex
(See full trace by running task with --trace)
```
---
### Author:  [@skriesch](https://meta.discourse.org/u/skriesch)
#### Post date: [13 Fevereiro , 2017 12:08 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/6 "2017-02-13T12:08:37Z")
The backup is giving the same output like in [https://meta.discourse.org/t/repairing-a-corrupted-database/16486/2](https://meta.discourse.org/t/repairing-a-corrupted-database/16486/2) now.
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [13 Fevereiro , 2017 13:00 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/7 "2017-02-13T13:00:10Z")
You are going to need to do a full pg dump / restore to recover from this
---
### Author:  [@skriesch](https://meta.discourse.org/u/skriesch)
#### Post date: [13 Fevereiro , 2017 13:29 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/8 "2017-02-13T13:29:18Z")
I tried it and get the same error like in the linked issue above:
```
su - postgres -c "pg_dumpall --database=discourse -f /shared/backups/default/discourse-2017-02-12.sql"
pg_dump: Dumping the contents of table "posts" failed: PQgetResult() failed.
pg_dump: Error message from server: ERROR: missing chunk number 0 for toast value 69836 in pg_toast_16766
pg_dump: The command was: COPY public.posts (id, user_id, topic_id, post_number, raw, cooked, created_at, updated_at, reply_to_post_number, reply_count, quote_count, deleted_at, off_topic_count, like_count, incoming_link_count, bookmark_count, avg_time, score, reads, post_type, vote_count, sort_order, last_editor_id, hidden, hidden_reason_id, notify_moderators_count, spam_count, illegal_count, inappropriate_count, last_version_at, user_deleted, reply_to_user_id, percent_rank, notify_user_count, like_score, deleted_by_id, edit_reason, word_count, version, cook_method, wiki, baked_at, baked_version, hidden_at, self_edits, reply_quoted, via_email, raw_email, public_version, action_code, image_url) TO stdout;
pg_dumpall: pg_dump failed on database "discourse", exiting
```
What does Zach Alexander mean with:
> Update: it was a problem with the cooked column of one particular post. Found it by binary searching through posts (looking for the “missing chunk number” error) in psql (after ssh into container).
> To fix the corrupted row, I first copied the contents of raw to cooked, making the row usable again (though it would probably render badly in a browser). Then I went to the Rails console, found the post again, and called p.cooked(p.raw) and p.save. p.cooked is now an HTML version of p.raw, as I presume it should be.
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [13 Fevereiro , 2017 13:36 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/9 "2017-02-13T13:36:33Z")
Something in your `posts` table is broken beyond repair, it may be a column in a row, or multiple rows and multiple columns.
What you can try doing is finding which `id` is broken by selecting it:
```plaintext
psql discourse
% select * from posts where id < 1000
% select * from posts where id > 1000 and id < 2000
```
and so on until you isolate the problem row.
Then if you are lucky you can just fix the row by deleting data from it, if you are unlucky you can just delete the row completely
---
### Author:  [@skriesch](https://meta.discourse.org/u/skriesch)
#### Post date: [13 Fevereiro , 2017 14:12 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/10 "2017-02-13T14:12:53Z")
```
postgres@xxx:~$ for ((i=0; i<13467; i++ )); do psql discourse -c "SELECT * FROM posts LIMIT 1 offset $i" >/dev/null || echo $i; done
ERROR: missing chunk number 0 for toast value 69836 in pg_toast_16766
1806
ERROR: unexpected chunk number 1 (expected 0) for toast value 69776 in pg_toast_16766
2725
ERROR: missing chunk number 0 for toast value 69797 in pg_toast_16766
2772
ERROR: missing chunk number 0 for toast value 69832 in pg_toast_16766
3032
ERROR: unexpected chunk number 1 (expected 0) for toast value 69853 in pg_toast_16766
3492
discourse=# select id from posts limit 1 offset 1806;
id
------
5987
(1 row)
discourse=# select id from posts limit 1 offset 2725;
id
------
5978
(1 row)
discourse=# select id from posts limit 1 offset 2772;
id
------
5982
(1 row)
discourse=# select id from posts limit 1 offset 3032;
id
------
5986
(1 row)
discourse=# select id from posts limit 1 offset 3492;
id
------
5990
(1 row)
discourse=# delete from posts where id in (5987, 5978, 5982, 5986, 5990);
DELETE 5
discourse=# VACUUM FULL posts;
VACUUM
```
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [13 Fevereiro , 2017 14:18 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/12 "2017-02-13T14:18:26Z")
Is everything working now?
---
### Author:  [@skriesch](https://meta.discourse.org/u/skriesch)
#### Post date: [13 Fevereiro , 2017 14:23 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/13 "2017-02-13T14:23:48Z")
Yes. Thanks!
Reindex was running and I can create backups again. 🙂
---
### Author:  [@system](https://meta.discourse.org/u/system)
#### Post date: [1 Maio , 2019 20:45 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/14 "2019-05-01T20:45:43Z")
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [2 Maio , 2019 03:19 UTC](https://meta.discourse.org/t/repairing-damaged-post-search-data-unable-to-create-database-dumps-and-execute-search-reindex/57046/15 "2019-05-02T03:19:23Z")